The Wheat from the Chaff #14

As part of my job, every week I read a lot of content about data, technology, and artificial intelligence (also known as AI). I filter them and send you the best ones, every Sunday.

Hello everyone 👋 ,

This week, we will address the theme of scenario planning and uncertainty. Our guests recently published an article in which they claim that ChatGPT and GPT-4 were trained using copyrighted books. The genres that top the list are science fiction and fantasy.

It's time to separate THE WHEAT FROM THE CHAFF…

Less is More

Humanity has always had the desire to predict the future. Examples range from ancient practices and traditions, such as dream interpretation, to more modern approaches, like forecasting market trends. In the business world, this is no different. The ability to develop plausible future scenarios represents a significant competitive advantage.

This introduction highlights the importance of assessing future scenarios in any aspect of our lives. Often, we assume that the future will be our present with a few (or no) changes. Two classic examples are our strategic plans and year-end resolutions, which typically focus on the following year. This mindset assumes that there will always be time or opportunity to take action.

The short-term concerns often overshadow our vision. Uncertainty is a structural characteristic of most ecosystems we are part of, including the business ecosystem. Accepting the existence of discontinuities and incorporating uncertainty as part of our mental model can assist in how we think and plan for the future. As for how to do this, we mentioned at the beginning of the text that predicting the future is an ancient desire, and as expected, many people have dedicated themselves to studying this subject and have developed interesting methods.

One of these individuals is Pierre Wack, who led Shell's scenario planning unit. According to Wack himself, this allowed the company to anticipate and position itself in the face of the 1973 oil crisis. In his 1985 article titled "Scenarios: Uncharted Waters Ahead," Wack argues that many analyses of the future are based on the idea that tomorrow will be like today. The article presents a logical construction based on real cases, and recommended reading. The text brings a still relevant observation in today's world: "In the next edition, I will discuss how we adapt the technique to develop short-term scenarios. As the time interval between decisions became shorter and shorter, this refinement proved to be necessary."

Peter Schwartz, who worked at Shell under the guidance of Pierre Wack, wrote the book "The Art of the Long View." Schwartz emphasizes that scenarios should not be used to select a preferred future and, even less, to work on creating that future. Instead, they serve to assist in making strategic decisions by considering various possible futures.

To imagine all these futures, Brian David Johnson created the technique of "science fiction prototyping" while working at Intel. This predictive technique uses elements of science fiction in building narratives within a 10-year window, enabling exploration of possibilities that do not yet exist and helping to visualize futures different from the present. The technique is structured in five steps:

1. Build a world, creating detailed environments and human characters based on your research;

2. Next, introduce a turning point, which can be a threat or a new technology;

3. This leads to discovering how the turning point affects your world, creating various outcomes;

4. After that, there needs to be another turning point: the solution (or the impossibility of solving the problem);

5. Finally, it is important to extract lessons from the prototyping process, including conversations about the story and its outcomes.

In building this fictional world, make use of scientific facts and data in order to create a sense of realism and plausibility in the story, giving the sensation that the story could happen at some future point.

You might be wondering if you're in the right newsletter, right? After all, today's content seems more related to futurology than artificial intelligence. However, if the next time you analyze scenarios you consider the current stage of AI as a turning point, I will have achieved my goal. The era of AI is just beginning!

Interesting topics in Brazil:

• 🔗 Link: President of the Senate presents a project to establish a regulatory framework for the use of AI in the country.

• 🔗 Link: Why is AI the topic of the moment?

Trending topics Global:

• 🔗 Link: Top 6 cybersecurity trends to keep an eye on in 2023

• 🔗 Link: How to avoid bias in fraud detection by AI

• 🔗 Link: Data Science vs. Machine Learning: What's the Difference?

Guest of the Week: Chang, Cramer, Soni, and Bamman

This week's guests are four researchers from the University of California, Berkeley. They recently published an article in which they state,

"We found that OpenAI models have memorized a wide range of copyrighted materials, and the degree of memorization is associated with how frequently snippets from these books appear on the web."

The authors emphasize in the article that there are profound ethical and reproducibility concerns. According to the Financial Times, the European Union is drafting new regulations related to AI. According to sources close to the matter, the forthcoming regulation may require companies like OpenAI (developer of ChatGPT) to disclose their use of copyrighted material in the training of their AI technologies.

Last month, Elon Musk threatened to sue Microsoft, accusing the software giant of illegally using Twitter data to train its AI model.

"They trained illegally using Twitter data," Musk tweeted. "Time to sue."

We would love your help sharing this website!

If you know someone that might like our posts, don’t be shy and share it with them! By the way, we appreciate this!

See you Sunday…

We always arrive in your inbox around 12:08 PM. Some email servers are stubborn and delayed... Others are even worse and throw us into spam and/or promotions. What's the logic behind this timing? ...None :)